Infrared thermography has demonstrated great potential in disease screening and detection, with applications in vascular disorder diagnostics, breast cancer detection, fever screening and dermatology. Infrared thermography is especially advantageous in skin cancer detection owing to its non-invasiveness and the ability to obtain measured data almost instantaneously. Despite these promises, the use of infrared thermography in skin cancer detection is limited to clinical trials. This may be due to the difficulty in correctly characterizing and classifying the thermal images into cancerous and non-cancerous lesions. This difficulty stems from the multiple factors that affect the heat signature across the human skin that subsequently influence the classification process. These factors include the presence of unaccounted local thermoregulation, variation in local thermal properties of the lesion and the presence of noise due to external factors, such as imperfection of the skin surface and ambient conditions.
It is hypothesised that the classification process of skin cancer based on its heat signature can be improved by using thermal images obtained from computational simulations of a dynamic thermal imaging procedure. Hence, this research project aims to investigate the feasibility of in silico dynamic thermography as a tool to elucidate the heat transfer mechanisms inside the tissue in the presence of skin abnormality. To achieve this, a human skin digital twin (HSDT) will be developed to simulate the thermophysiology of the actual human skin through the appropriate selection of bioheat transfer models that best describe the thermophysical processes of skin under basal and pathological conditions. In silico dynamic thermography will be carried out on the HSDT, where the computationally generated thermal images will be treated as ‘clinical data’ for segmentation and feature extraction.
There are many advantages in using computational models to generate the required synthetic thermal images for these segmentation and feature extraction processes. Firstly, computational models allow the flexibility to generate virtually infinite number of test cases that mimic a variety of different skin cancer thermophysiology. Secondly, external factors can be incorporated into the model to generate varying degrees of noise to assist with the delineation of the main signal from the noise during the classification process. Thirdly, synthetic data opens the possibility for the identification of new potential biomarkers for skin cancer detection that may not be possible using actual thermal images. This can help to elevate the versatility of infrared thermography as an effective non-invasive skin cancer detection tool.
+ This position is open to both Malaysians and non-Malaysians.
+ A First Class Bachelor degree in a relevant area of engineering, e.g. Mechanical, Biomedical and Chemical Engineering)
+ Demonstrated strong knowledge in the field of advanced solid mechanics, applied mathematics and computational modelling
+ Experience with finite element software, such as COMSOL Multiphysics is an added advantage
+ Candidates with prior research experience and with good publication record will have added advantages.
+ Independent and with good communication skills
+ Interested candidates can send a complete CV, copies of academic transcripts and/or other relevant documents to Dr Ean Hin Ooi at email@example.com
+ In taking this project, you will join the Biomedical Engineering Modelling and Simulation (BEMS) group led by Dr Ean Hin Ooi of the School of Engineering, Monash University Malaysia.
+ Only shortlisted candidates will be notified.
+ This position will commence only in Year 2023.